I'm conflicted about vau: on the one hand, it's the cleanest expression of the "Lisp ideal" I've seen yet. Replacing special forms with fexprs makes everything

I'm conflicted about vau: on the one hand, it's the cleanest expression of the "Lisp ideal" I've seen yet. Replacing special forms with fexprs makes everything "first class", and lets the programmer mold the language any way they see fit, without being boxed in by someone else's opinion of what computing should look like. On the other hand, it really spits in the face of any kind of static analysis. Assigning types to vau based programs is going to be freakishly difficult if not impossible in practice. Some clever form of dynamic analysis/partial evaluation is going to be needed if you want to regain the benefits of static languages, like type checking and compilation. That's another PhD thesis or two, though...